Ji-Hoon Nam
Ji-Hoon Nam has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Byung Woo Cho
Department of Orthopaedic Surgery, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea
3 shared publications - Jun Young Park
Department of Orthopaedic Surgery, Yongin Severance Hospital, Yonsei University College of Medicine, Yongin-si, Republic of Korea
3 shared publications - Kwan Kyu Park
Department of Orthopaedic Surgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea
3 shared publications - Kyoung-Tak Kang
Department of Mechanical Engineering, Yonsei University, Seoul, Republic of Korea
3 shared publications - Hyuck Min Kwon
Department of Orthopaedic Surgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea
2 shared publications - Jong-Keun Kim
Department of Pharmacology, Chonnam National University Medical School, Hwasun, Korea
2 shared publications
Latest Publications
- Open Access
- 26 May 2026
- Published by British Editorial Society of Bone & Joint Surgery in Bone & Joint Research
- Vol. 15 (5) , 566-576
Does tibial sclerosis severity influence bone resorption risk following total knee arthroplasty?
Aims: Stress shielding, particularly affecting the proximal medial tibial bone, is a significant concern following total knee arthroplasty (TKA). This study investigates how tibial sclerosis severity, implant configurations, and rotational alignment techniques influence bone resorption patterns in the medial proximal tibia. Methods: Finite element analysis (FEA) was performed using tibial models with four sclerosis levels (0 to 3) representing progressive sclerosis penetration. Four implant configurations combining baseplate shapes (symmetrical vs anatomical) and keel designs (longer (40 mm) vs shorter (30 mm)) were evaluated with two rotational alignment protocols, generating 32 models (4 × 4 × 2). Bone resorption rates in four medial cortical regions were quantified using strain energy density (SED) criteria. Results: A dose-dependent relationship was demonstrated between sclerosis severity and bone resorption rates. In the total medial cortical region, a substantial increase occurred at Level 3 over Level 0 (67.8% vs. 64.8%, respectively, p < 0.001). Keel design emerged as the dominant factor, with shorter keels demonstrating significantly lower absolute resorption rates than longer keels across all sclerosis levels (59.9% to 64.0% vs. 69.8% to 71.7%, p = 0.029). However, when changes were analyzed relative to baseline sclerosis severity, shorter keels exhibited significantly greater increases in bone resorption rates than longer keels (4.1% vs 1.9% at Level 3, p = 0.029), indicating greater sensitivity to sclerosis progression. Contrary to implant configurations, no significant differences were observed between the maximum-coverage method and Insall’s method across all sclerosis severity levels (p = 0.486). Conclusion: Proximal medial tibial sclerosis severity demonstrates a clear dose-dependent relationship with bone resorption risk following TKA. Although shorter keel configurations provide lower absolute bone resorption rates, they exhibit greater sensitivity to sclerosis progression than longer keels. Rotational alignment techniques showed no significant effects on bone resorption patterns in sclerotic bone conditions. These findings support preoperative sclerosis assessment for surgical planning and highlight the complex interaction between implant design and bone quality. Cite this article: Bone Joint Res 2026;15(5):566–576.
- 26 May 2026
- Published by British Editorial Society of Bone & Joint Surgery in Bone & Joint Research
- Vol. 15 (5) , 566-576
Open AccessDoes tibial sclerosis severity influence bone resorption risk following total knee arthroplasty?
Aims: Stress shielding, particularly affecting the proximal medial tibial bone, is a significant concern following total knee arthroplasty (TKA). This study investigates how tibial sclerosis severity, implant configurations, and rotational alignment techniques influence bone resorption patterns in the medial proximal tibia. Methods: Finite element analysis (FEA) was performed using tibial models with four sclerosis levels (0 to 3) representing progressive sclerosis penetration. Four implant configurations combining baseplate shapes (symmetrical vs anatomical) and keel designs (longer (40 mm) vs shorter (30 mm)) were evaluated with two rotational alignment protocols, generating 32 models (4 × 4 × 2). Bone resorption rates in four medial cortical regions were quantified using strain energy density (SED) criteria. Results: A dose-dependent relationship was demonstrated between sclerosis severity and bone resorption rates. In the total medial cortical region, a substantial increase occurred at Level 3 over Level 0 (67.8% vs. 64.8%, respectively, p < 0.001). Keel design emerged as the dominant factor, with shorter keels demonstrating significantly lower absolute resorption rates than longer keels across all sclerosis levels (59.9% to 64.0% vs. 69.8% to 71.7%, p = 0.029). However, when changes were analyzed relative to baseline sclerosis severity, shorter keels exhibited significantly greater increases in bone resorption rates than longer keels (4.1% vs 1.9% at Level 3, p = 0.029), indicating greater sensitivity to sclerosis progression. Contrary to implant configurations, no significant differences were observed between the maximum-coverage method and Insall’s method across all sclerosis severity levels (p = 0.486). Conclusion: Proximal medial tibial sclerosis severity demonstrates a clear dose-dependent relationship with bone resorption risk following TKA. Although shorter keel configurations provide lower absolute bone resorption rates, they exhibit greater sensitivity to sclerosis progression than longer keels. Rotational alignment techniques showed no significant effects on bone resorption patterns in sclerotic bone conditions. These findings support preoperative sclerosis assessment for surgical planning and highlight the complex interaction between implant design and bone quality. Cite this article: Bone Joint Res 2026;15(5):566–576.
- Open Access
- 2 March 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14, e86155
Enhanced Magnetic Resonance Imaging–Based Knee Cartilage Segmentation Using a Swin-UNet Conditional Generative Adversarial Network: Development and Validation Study
Background: Accurate segmentation of cartilage from magnetic resonance imaging (MRI) is crucial for the diagnosis and surgical planning of knee osteoarthritis. However, manual segmentation is time-consuming, and conventional computed tomography–based surgical systems are limited by their inability to visualize cartilage. Objective: This study aimed to develop a clinically targeted deep learning framework, the Swin-UNet conditional generative adversarial network (cGAN), for the automatic segmentation of femoral and tibial cartilage in MRI. We then evaluated its performance against conventional UNet, UNet cGAN, and Swin-UNet baseline models. Methods: Our dataset comprised 232 knee MRI scans. We conducted quantitative experiments on the proposed Swin-UNet cGAN model and compared the results with those of widely used UNet, UNet cGAN, and Swin-UNet models for femoral and tibial cartilage segmentation, using the Dice similarity coefficient, mean intersection over union, 95th percentile Hausdorff distance, and average symmetric surface distance. All performance metrics were statistically analyzed. In addition, the performance of the Swin-UNet cGAN model was evaluated on an external validation dataset. Results: The proposed Swin-UNet cGAN achieved the highest mean Dice similarity coefficient and intersection over union scores for both femoral and tibial cartilage segmentation, demonstrating performance statistically comparable to the best-performing baseline (UNet) in the tibia. Regarding distance metrics (average symmetric surface distance and 95th percentile Hausdorff distance), the proposed model significantly outperformed all baselines in the tibia while achieving results comparable to the UNet cGAN in the femur. It also maintained consistently high segmentation performance on both the internal test set and an external validation dataset. Conclusions: These findings indicate that the proposed Swin-UNet cGAN achieves more accurate knee cartilage segmentation than UNet, UNet cGAN, and Swin-UNet, particularly in terms of boundary accuracy, while maintaining promising generalizability performance across both internal testing and external validation cohorts. This MRI-based deep learning approach addresses critical limitations of computed tomography–based patient-specific instrumentation systems by providing cartilage visualization, potentially improving surgical precision and outcomes in total knee arthroplasty.
- 2 March 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14, e86155
Open AccessEnhanced Magnetic Resonance Imaging–Based Knee Cartilage Segmentation Using a Swin-UNet Conditional Generative Adversarial Network: Development and Validation Study
Background: Accurate segmentation of cartilage from magnetic resonance imaging (MRI) is crucial for the diagnosis and surgical planning of knee osteoarthritis. However, manual segmentation is time-consuming, and conventional computed tomography–based surgical systems are limited by their inability to visualize cartilage. Objective: This study aimed to develop a clinically targeted deep learning framework, the Swin-UNet conditional generative adversarial network (cGAN), for the automatic segmentation of femoral and tibial cartilage in MRI. We then evaluated its performance against conventional UNet, UNet cGAN, and Swin-UNet baseline models. Methods: Our dataset comprised 232 knee MRI scans. We conducted quantitative experiments on the proposed Swin-UNet cGAN model and compared the results with those of widely used UNet, UNet cGAN, and Swin-UNet models for femoral and tibial cartilage segmentation, using the Dice similarity coefficient, mean intersection over union, 95th percentile Hausdorff distance, and average symmetric surface distance. All performance metrics were statistically analyzed. In addition, the performance of the Swin-UNet cGAN model was evaluated on an external validation dataset. Results: The proposed Swin-UNet cGAN achieved the highest mean Dice similarity coefficient and intersection over union scores for both femoral and tibial cartilage segmentation, demonstrating performance statistically comparable to the best-performing baseline (UNet) in the tibia. Regarding distance metrics (average symmetric surface distance and 95th percentile Hausdorff distance), the proposed model significantly outperformed all baselines in the tibia while achieving results comparable to the UNet cGAN in the femur. It also maintained consistently high segmentation performance on both the internal test set and an external validation dataset. Conclusions: These findings indicate that the proposed Swin-UNet cGAN achieves more accurate knee cartilage segmentation than UNet, UNet cGAN, and Swin-UNet, particularly in terms of boundary accuracy, while maintaining promising generalizability performance across both internal testing and external validation cohorts. This MRI-based deep learning approach addresses critical limitations of computed tomography–based patient-specific instrumentation systems by providing cartilage visualization, potentially improving surgical precision and outcomes in total knee arthroplasty.
- Open Access
- 20 October 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Enhanced Magnetic Resonance Imaging–Based Knee Cartilage Segmentation Using a Swin-UNet Conditional Generative Adversarial Network: Development and Validation Study (Preprint)
BACKGROUND Accurate segmentation of cartilage from magnetic resonance imaging (MRI) is crucial for the diagnosis and surgical planning of knee osteoarthritis. However, manual segmentation is time-consuming, and conventional computed tomography–based surgical systems are limited by their inability to visualize cartilage. OBJECTIVE This study aimed to develop a clinically targeted deep learning framework, the Swin-UNet conditional generative adversarial network (cGAN), for the automatic segmentation of femoral and tibial cartilage in MRI. We then evaluated its performance against conventional UNet, UNet cGAN, and Swin-UNet baseline models. METHODS Our dataset comprised 232 knee MRI scans. We conducted quantitative experiments on the proposed Swin-UNet cGAN model and compared the results with those of widely used UNet, UNet cGAN, and Swin-UNet models for femoral and tibial cartilage segmentation, using the Dice similarity coefficient, mean intersection over union, 95th percentile Hausdorff distance, and average symmetric surface distance. All performance metrics were statistically analyzed. In addition, the performance of the Swin-UNet cGAN model was evaluated on an external validation dataset. RESULTS The proposed Swin-UNet cGAN achieved the highest mean Dice similarity coefficient and intersection over union scores for both femoral and tibial cartilage segmentation, demonstrating performance statistically comparable to the best-performing baseline (UNet) in the tibia. Regarding distance metrics (average symmetric surface distance and 95th percentile Hausdorff distance), the proposed model significantly outperformed all baselines in the tibia while achieving results comparable to the UNet cGAN in the femur. It also maintained consistently high segmentation performance on both the internal test set and an external validation dataset. CONCLUSIONS These findings indicate that the proposed Swin-UNet cGAN achieves more accurate knee cartilage segmentation than UNet, UNet cGAN, and Swin-UNet, particularly in terms of boundary accuracy, while maintaining promising generalizability performance across both internal testing and external validation cohorts. This MRI-based deep learning approach addresses critical limitations of computed tomography–based patient-specific instrumentation systems by providing cartilage visualization, potentially improving surgical precision and outcomes in total knee arthroplasty.
- 20 October 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessEnhanced Magnetic Resonance Imaging–Based Knee Cartilage Segmentation Using a Swin-UNet Conditional Generative Adversarial Network: Development and Validation Study (Preprint)
BACKGROUND Accurate segmentation of cartilage from magnetic resonance imaging (MRI) is crucial for the diagnosis and surgical planning of knee osteoarthritis. However, manual segmentation is time-consuming, and conventional computed tomography–based surgical systems are limited by their inability to visualize cartilage. OBJECTIVE This study aimed to develop a clinically targeted deep learning framework, the Swin-UNet conditional generative adversarial network (cGAN), for the automatic segmentation of femoral and tibial cartilage in MRI. We then evaluated its performance against conventional UNet, UNet cGAN, and Swin-UNet baseline models. METHODS Our dataset comprised 232 knee MRI scans. We conducted quantitative experiments on the proposed Swin-UNet cGAN model and compared the results with those of widely used UNet, UNet cGAN, and Swin-UNet models for femoral and tibial cartilage segmentation, using the Dice similarity coefficient, mean intersection over union, 95th percentile Hausdorff distance, and average symmetric surface distance. All performance metrics were statistically analyzed. In addition, the performance of the Swin-UNet cGAN model was evaluated on an external validation dataset. RESULTS The proposed Swin-UNet cGAN achieved the highest mean Dice similarity coefficient and intersection over union scores for both femoral and tibial cartilage segmentation, demonstrating performance statistically comparable to the best-performing baseline (UNet) in the tibia. Regarding distance metrics (average symmetric surface distance and 95th percentile Hausdorff distance), the proposed model significantly outperformed all baselines in the tibia while achieving results comparable to the UNet cGAN in the femur. It also maintained consistently high segmentation performance on both the internal test set and an external validation dataset. CONCLUSIONS These findings indicate that the proposed Swin-UNet cGAN achieves more accurate knee cartilage segmentation than UNet, UNet cGAN, and Swin-UNet, particularly in terms of boundary accuracy, while maintaining promising generalizability performance across both internal testing and external validation cohorts. This MRI-based deep learning approach addresses critical limitations of computed tomography–based patient-specific instrumentation systems by providing cartilage visualization, potentially improving surgical precision and outcomes in total knee arthroplasty.